Now I go and create a painting called Sunflowers. Does Van Gogh's estate own some of my work?
Should every HN comment I write include a full list of everything I've ever read? What about a commercial work like a book, should that include a list of everything I've read or heard in the past 35 years of my life? That's a lot of attribution to keep track of ...
edit: I guess my question is where does derivativity end and creativity begin?
What Stable Diffusion does is effectively say "it doesn't matter how this was made" for every single artifact it creates, providing neither attribution nor acknowledgement. This happens even when it's clearly been influenced by, and in some cases is directly copying, elements or a whole of a very small set of highly influential inputs.
If a human author did that, people would probably be upset, if not outright accuse them of plagiarism.
But the controversy over Stable Diffusion is whether they owe the human artists a cut or not, right?
> What they _are_ saying is not to claim that you didn't train on Van Gogh
Stable Diffusion is definitely not claiming that it didn't train on human artists.
The controversy of Stable Diffusion is attribution. I could not care less about receiving a cut. Rather, I want my life's work in pushing particular style further than anyone has before to be recognized primarily so the next artist can trace the lineage and improve upon it.
> Stable Diffusion is definitely not claiming that it didn't train on human artists.
Stable diffusion claims that attribution is not important because "all artists copy each other".
Have you studied all masters and eclipsed them, without acknowledging them so they can share in your fame? Will you use your untiring, undying skill to flood the market with your work so "artist" ceases to be a profession?
That's categorically new, and worth preventing.
This would be bad for present-day artists who've invested their lives in developing the skills of putting brush strokes on the page; those skills would become worthless. But the flip side is that Stable Diffusion lets anyone make art easily, generating whatever image they can dream up, without having to spend years developing the skills first.
So even though Stable Diffusion is bad for present-day _artists_, I suspect it's actually good for _art_ in the long term. "Anyone can easily create an endless supply of art, based on any idea they can dream up" seems like a good thing to me.
Put another way, it is the belief that the 2020s are the peak of manual human art.
Imagine if we'd felt that way a thousand years ago, or even a hundred. Is it worth being wrong?
Also, I think we have to weigh that tradeoff against all the new things that will be created if art becomes accessible to 100x as many people. On net, I think Stable Diffusion opens many more doors for art than it closes.
Its not simply that he had good technique. Its that he felt something deep down. He felt it ever since he was a child, and made constant and frustrating (to him) attempts to then manifest that feeling. He would sometimes linger by a roadside too long and become his own little philosopher, seeing a new treatise in the ebb and flow of landscape from the sky. He had loves and heartbreak and fantasies, and found no other place to recover from those things than in patient work with the canvas. His whole life was a sacrifice, a sacred devotion to the spirit of humanity only art can capture so thoroughly and universally. You can't just do that.
I'm just saying, a guy like Van Gogh comes around just a few times a century, he is not somebody you can simply encode into a lot of different matrices, it takes a certain human life to make art like him, or even more simply, to simply see the world like him. Its not just about the way the sunflowers look, its about the choice in the first place to paint them, and making that choice in the particular context he was in, at the particular time in history he did. Its important that it was this painting that was painted after all the other paintings that had come before, both in particular for Van Gogh and for the world. And you can't ever recreate it, because its already done, and you can never step in the same river twice.
In general, and no offense to you in particular, but you will be forgotten as an artist. If this wasn't a certainty before, it is one now with all art generation and such. But don't take that as a bad thing, it is more than anything liberating. Try to remove all expectation and ego from what you are trying to make, and that will bring you closer to Van Gogh's work than any particular technique he used--your artwork should be completely personal, entirely internal, right up to the point that someone else see's it for the first time. It should only be judged on a sui generis rubric, more like a dream than a product. Stop, in general, trying to prove yourself on the basis of the tools you use. You could use sticks-and-twine or a million-GPU-DALL-E thing, whatever, its all orthogonal to the effect the final product might have, and that effect itself is only influenced by your attitude, your sincerity, and your vulnerability, as it is applied to the work.
(That is my maybe my main issue with the AI art bots, do the artists who work with them even feel vulnerable about their work? Can artistic sincerity and enthusiasm exist in AI art? Do AI artists feel the literally metaphysical stakes of what they are doing like Van Gogh did? Or is it all so constantly folded back into the idea of AI art itself? Is it all just different campaign posters for the cause of legitimizing itself as "real" art? Maybe just some time needs to pass, but I hope artists can still get nervous in the future, can still struggle at all to bring something out. I worry that people using this technology are too worried about proving themselves, they are not opening themselves up the possibility of failure that all artists need. They are not in a silent enough room to be able to hear what there brain is actually saying, only trying to respond respond respond. It is usually the struggle the artist has with themselves that is transmuted into the work such that it makes us cry or feel profound joy or sadness or whatever, we still need that psychic energy in the world, it doesn't need to be solved.)
To be like Van Gogh, you need to make art as if you would die and go to literal hell if you didn't, or go there if you made the "wrong" art. If you feel something like that, you are on the right track, but you still have 0.001% chance of dealing with that illness as productively as Van Gogh did.
The next Van Gogh, or the next Francis Bacon, or Rothko, or Cezanne, or whoever, will probably not have the machine-capturable-"style" of any of those people, but she will share in the particular mental complex they had, the same spiritual curse which drives one to reach past a scientific or practical world, to tap into something pure, almost completely formal, that cannot simply be taught or transmitted, much less encoded.
Sometimes, yes. John Q Nerd playing around with it might not, but the "good AI artists" I follow on Twitter certainly do.
> Can artistic sincerity and enthusiasm exist in AI art?
Unequivocally yes.
> Do AI artists feel the literally metaphysical stakes of what they are doing like Van Gogh did?
Some of them seem to, though I dunno precisely what Van Gogh felt so I guess I'm not sure.
> Is it all just different campaign posters for the cause of legitimizing itself as "real" art?
Not sure where you find this stuff, but it sounds really boring. Certainly not all of it is.
Your unsolicited advice falls flat (to me), sorry. My unsolicited advice to you would be to psychologize less.
But I really really do wish you the best with this stuff, the future is bright and interesting either way for art!
Otherwise, it depends what the sunflowers you painted look like. You’ve seen the work that it is alleged you copied, so stuff that looks a lot like copying is assumed to be copying, roughly speaking. If you produced a work that reproduces substantial elements that were original in Van Gogh’s work, then Van Gogh has a case in copyright infringement. That can include the overall composition of the painting, some particular brushwork, etc. Techniques are not protected by copyright (that is generally the domain of patents, but afaik there are no extant patents on how to use a brush), so you can copy the techniques as much as you like, as long as you are not copying actual brushstrokes to reproduce the painting.
Copyright is concerned fundamentally with the degree to which you are stealing the market for the original work. It was created to allow authors etc to profit off their hard work, rather than labour for years writing a book only for someone else to make all the money from selling it. The existence of copyright makes it possible to make money despite the fact that copying someone else’s work has always been easier and cheaper than making something new. You could spend ten years writing a book, and it would take someone a day to print a copy, and twenty other printers could do it and none of them would pay you… so why would you write a book in the first place unless it were a political/religious pamphlet? To guide your interpretation of copyright, think about the degree to which you are stealing a market for Van Gogh paintings. If you produce a very close replica, then yeah, you are stealing some of the market. Photographing the painting and selling prints does the same thing. Producing something close from scratch without reference to the original is absolutely fine, if there is no copying then there is no infringement. The copyright statute just doesn’t want you stealing Van Gogh’s market for his own works using the crutch of copying his stuff, and through no original effort of your own.
Obviously AI art is a huge problem for this fundamental goal, because it hides the copying that it does. Someone selling a copy of your book without a license is easy to identify. AI art lets you take some art you like, launder it a bit, and then use it in your marketing materials without the original artists being able to notice the infringement and sue you for it. Fundamentally it is like money laundering, the same way AI data analysis has often been used for laundering human biases. That’s a huge problem and really subverts the entire intention of copyright. It steals the market and hides it at the same time.
You can also see how using an AI tool to launder images is just copying with a small amount of original human labour curating the input works. You can have some credit for your labour picking them, but you can’t really pretend the source images’ copyrights just disappear. The only issue is which humans are owed royalties. There’s a lot of debate going on about whether computers doing art is just like painters studying, just like you said, but if you analyse it as the laws do only in terms of the humans involved, you have a bunch of artists, and someone operating a lever labelled “mix these paintings so nobody can recognise my IP theft” that will vastly outpace their production and steal 99% of the market very quickly. It’s the invention of the printing press all over again.
If the copyright laws are too difficult to apply (eg an AI painting is such that it has laundered away any ability to pick out which bits infringe a human’s original painting used as input) then rather than everyone throwing up their hands and admitting defeat, the laws will probably change to better serve their original purpose. For example, AI paintings may have to pay royalties to every single artist used as input regardless of whether individually identifiable instances of copying are visible in the output.